Recognises a criterion grading two skills

T-201 · Skill knobit · Activity: Write and review test items · Competence: Skill assessment

One criterion, one score, two knobits that can fail independently.

Can AI do this?

AI does this well Most of the tested AI models master this. 3 of 4 models in the current benchmark master it.

Model	Tier	Probability of mastery	Status
gemini-4-argon	Flagship	98%	Mastered
gemini-3.5-flash-lite	Low cost	68%	Not shown
gpt-6-luna	Low cost	98%	Mastered
claude-haiku-5-5	Low cost	98%	Mastered

Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.

What it looks like in practice

Expected result: The criterion split into parts with an item each

How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: C1 (2 points) "Finds the duplicate customers and writes the merge rule" Acceptable: (a) finds the duplicates, (b) writes the merge rule, 1 point each, each part its own item

Also known as: Criterion grading two skills

Related knobits

Builds on

T-251 Explains rubric-based test item design

Usually followed by

T-231 Writes a rubric criterion observing one knobit set

All 15 knobits of Write and review test items.

Source: Review checklist 02-test-items, row 1 (KST; Practice)